The feasibility of using a protocol (adaptive or randomized) to govern the process of selecting a data rate for a two-rate communication system subject to jamming is explored. Two quantities, average efficiency for independent jamming and worst case efficiency for intelligent jamming, are defined as performance measures of an adaptive protocol. A particular fatal property, called zero looping, of an adaptive protocol is discussed. One simple method for eliminating zero loops of adaptive protocols is given. Under certain jamming strategy constraints, the average throughput of a randomized protocol as well as the average efficiency of an adaptive protocol are computed and compared.
|State||Published - 1 Dec 1986|